Project 8 - Decision trees and neural networks

Project 8 - Decision trees and neural networks
Due: Fri Apr 5, 2024 11:30am
Ungraded, 100 Possible Points
In Progress
NEXT UP: Submit Assignment
Unlimited Attempts Allowed
Available until Apr 5, 2024 11:30am

This is a short, exploratory project. You will find a dataset and apply both decision trees and neural networks to it for classification. You don't need to use the same explanatory variables for both types of classification but it will be helpful if at least some of the variables overlap.

For the decision trees you should visualize at least two trees (using code we saw in class) at different maximum depths. Discuss the effect the depth has on bias and variance. Discuss the choices the algorithm makes in which attributes to split on. For example, at a given tree level do all nodes split on the same attribute? Why or why not?

For the neural networks, you should try at least three different architectures with varying numbers of hidden layers and nodes in each layer. Visualize each architecture. You can use code included in the lecture neural networks code for visualization or you can use a different visualization system (this one Links to an external site. is pretty awesome). We used sklearn's implementation of neural networks in the example code, but you are welcome to use other implementations (e.g. TensorFlow) if you choose. Discuss any correlations you see, or the lack thereof, between your edge weights and the decision trees.

Project rubric (4)
CriteriaRatingsPts
Project rubric (4)
CriteriaRatingsPts
Introduction
10 pts
Full Marks
0 pts
No Marks
/ 10 pts
Dataset
20 pts
Full Marks
0 pts
No Marks
/ 20 pts
Analysis technique
30 pts
Full Marks
0 pts
No Marks
/ 30 pts
Results
35 pts
Full Marks
0 pts
No Marks
/ 35 pts
Presentation Slides
5 pts
Full Marks
0 pts
No Marks
/ 5 pts
Total Points: 0
Keep in mind, this submission will count for everyone in your Project 8 group.

Choose a submission type

or